Poll Attempts an Unbiased Methodology

How to make a responsible survey?

For those at the Institute of Politics (IOP) it was more difficult than it looked.

The questions for the survey were originally created in brainstorming sessions among students at the IOP in the early fall.

These questions were given to Harvard Opinion Research Program Director Robert J. Blendon so that he could reformulate them into fair and unbiased survey questions. Blendon has served as a survey consultant for several professional papers, most notably the Washington Post's 1995 race poll.

"We in no way told the students what to ask, we told them to tell us what do you want to know and we will offer them a balanced survey," said Blendon, who also is Lee professor of health policy and management at the Harvard School of Public Health.

Blendon reviewed the questions on the survey after it was completed at the request of The Crimson and said he would vouch for all but the demographic categories.

The survey was conducted by telephone from The Crimson. The pollsters began by choosing a random place in a preliminary student phone book, and then called every tenth student. Students who were not home were called again.

Pollsters reached 530 students of the 650 they were trying to contact. Of the students reached, 30.7 percent were first-years, 23.7 were sophomores, 24.1 percent were juniors and 21.3 percent were seniors.

Racially, 59.7 percent identified themselves as white or Caucasian, 19.8 percent Asian-American, 11.3 percent African-American or black, 3.2 Mexican American or Chicano, 1.1 percent Puerto Rican, 4.5 percent other Latino or Hispanic and 0.4 percent Native American.

For tabulating purposes, categories three to five were aggregated to form a broader Latino group.

The questions, and many of the results, were also reviewed at the request of The Crimson by Associate Professor of Statistics Alan Zaslavsky, who teaches "Statistics 160: Survey Method."

"It looks to me like this is a good start to understanding race at Harvard," said Zaslavsky. "If you wanted to find out more, the questions would have to be more refined."

Both professors expressed some concern about the nature of the demographic categories. On census forms, Hispanic is given as a separate yes or no category, because Hispanic citizens can be of any race.

IOP members said they were aware of Blendon's objection but chose to take the categories directly from the Harvard admissions form.